Introduction: Cloud computing is considered a remarkable paradigm shift in Information Technology (IT), offering scalable and virtualized resources to end users at a low cost in terms of infrastructure and maintenance. These resources offer an exceptional degree of flexibility and adhere to established standards, formats, and networking protocols while being managed by several management entities. However, the existence of flaws and vulnerabilities in underlying technology and outdated protocols opens the door for malicious network attacks.Methods: This study addresses these vulnerabilities by introducing a method for classifying attacks in Infrastructure as a Service (IaaS) cloud environments, utilizing machine learning methodologies within a digital forensics framework. Various machine learning algorithms are employed to automatically identify and categorize cyber-attacks based on metrics related to process performance. The dataset is divided into three distinct categories—CPU usage, memory usage, and disk usage—to assess each category’s impact on the detection of attacks within cloud computing systems.Results: Decision Tree and Neural Network models are recommended for analyzing disk-related features due to their superior performance in detecting attacks with an accuracy of 90% and 87.9%, respectively. Neural Network is deemed more suitable for identifying CPU behavior, achieving an accuracy of 86.2%. For memory-related features, K-Nearest Neighbor (KNN) demonstrates the best False Negative Rate (FNR) value of 1.8%.Discussion: Our study highlights the significance of customizing the selection of classifiers based on the specific system feature and the intended focus of detection. By tailoring machine learning models to particular system features, the detection of malicious activities in IaaS cloud environments can be enhanced, offering practical insights into effective attack classification.

Phishing email assaults have been a prevalent cybercriminal tactic for many decades. Various detectors have been suggested over time that rely on textual information. However, to address the growing prevalence of phishing emails, more sophisticated techniques are required to use all aspects of emails to improve the detection capabilities of machine learning classifiers. This paper presents a novel approach to detecting phishing emails. The proposed methodology combines ensemble learning techniques with various variables, such as word frequency, the presence of specific keywords or phrases, and email length, to improve detection accuracy. We provide two approaches for the planned task; The first technique employs ensemble learning soft voting, while the second employs weighted ensemble learning. Both strategies use distinct machine learning algorithms to concurrently process the characteristics, reducing their complexity and enhancing the model’s performance. An extensive assessment and analysis are conducted, considering unique criteria designed to minimize biased and inaccurate findings. Our empirical experiments demonstrates that using ensemble learning to merge attributes in the evolution of phishing emails showcases the competitive performance of ensemble learning over other machine learning algorithms. This superiority is underscored by achieving an F1-score of 0.90 in the weighted ensemble method and 0.85 in the soft voting method, showcasing the effectiveness of this approach.

The hardware and software of a computer are controlled by its operating system (OS), which performs essential tasks such as input and output processing, file and memory management, and the management of peripheral devices such as disk drives and printers. Application software refers to programs designed for specific purposes, these applications, often freely available and open source, contribute to the rising number of downloads. In the third quarter of 2022, combined downloads from the Apple App Store and Google Play Reached an estimated 35.3 billion. However, the prevalence of insecurity in these applications and technologies heightens the potential for cybercrimes. Protection against unauthorized intruders is crucial in identifying malicious applications. Machine learning (ML) serves as a promising avenue for detecting malware attacks, offering potential solutions to bolster cybersecurity measures. We propose a novel approach utilizing ML to enhance malware detection accuracy by segmenting datasets into distinct groups. Our research employs supervised ML techniques on the CICMaldroid2020 dataset, which includes comprehensive information such as intent actions, permissions, and sensitive APIs. The dataset was partitioned into four groups, each containing 150 features, and analyzed across four experiments to distinguish between attack and benign classes. Our proposed model demonstrated exceptional performance, with the random forest algorithm achieving an accuracy of 98.6% and a precision of 98.75%. These results highlight the effectiveness of our segmentation approach and its significant contribution to advancing malware detection in Android applications, offering a promising direction for future cybersecurity solutions.

Electronic health records (EHRs) are widely used in healthcare institutions worldwide, containing vast amounts of unstructured textual data. However, the sensitive nature of Protected Health Information (PHI) embedded within these records presents significant privacy challenges, necessitating robust de-identification techniques. This paper introduces a novel approach, leveraging a Bi-LSTM-CRF model to achieve accurate and reliable PHI de-identification, using the i2b2 dataset sourced from Harvard University. Unlike prior studies that often unify Bi-LSTM and CRF layers, our approach focuses on the individual design, optimization, and hyperparameter tuning of both the Bi-LSTM and CRF components, allowing for precise model performance improvements. This rigorous approach to architectural design and hyperparameter tuning, often underexplored in the existing literature, significantly enhances the model’s capacity for accurate PHI tag detection while preserving the essential clinical context. Comprehensive evaluations are conducted across 23 PHI categories, as defined by HIPAA, ensuring thorough security across critical domains. The optimized model achieves exceptional performance metrics, with a precision of 99%, recall of 98%, and F1-score of 98%, underscoring its effectiveness in balancing recall and precision. By enabling the de-identification of medical records, this research strengthens patient confidentiality, promotes compliance with privacy regulations, and facilitates safe data sharing for research and analysis.

N this research work, six different filters are applied on a low resolution 8 b/pixel gray-scale images, which operate on small sub-images (windows of 3×3 to 11×11 pixels). The enhanced images are used to compare the efficiency of the different six filters using the peak signal to noise ratio (PSNR) image quality measure. Noise peak elimination filter (PSNR)=36.63) outperforms others, such as median filter (PSNR=36.61), while corruption estimation (PSNR=36.03) significantly cuts processing time by only processing the corrupted pixels while maintaining image details. Mean filter (PSNR=34.05) is sensitive to outliers, which cause the image's sharpness and fine features to be lost. By avoiding averaging across edges, bimodal-averaging filter (PSNR=35.30), which improves on the mean filter, chooses the mean of the biggest population. The median-mean filtering (PSNR=36.32), which combines median and mean filters and determines the output pixel by averaging the median and some nearby pixels, is another improvement above averaging.

The rapid proliferation of the internet and advancements in communication technologies have significantly improved networking and increased data vol ume. This phenomenon has subsequently caused a multitude of novel attacks, thereby presenting significant challenges for network security in the intrusion detection system (IDS). Moreover, the ongoing threat from authorized entities who try to carry out various types of attacks on the network is a concern that must be handled seriously. IDS are used to provide network availability, confidentiality, and integrity by employing machine learning (ML) and deep learning (DL) algorithms. This research aimed to study the impacts of the binary and multi-attack instances label by establishing IDS that leverages hybrid algorithms, including artificial neural networks (ANN), random forest (RF), and logistic model trees (LMTs). The paper addresses challenges such as data pre processing, feature selection, and managing imbalanced datasets by applying synthetic minority oversampling technique (SMOTE) and Pearson’s correlation methodologies. The IDS was tested using network security laboratory knowledge discovery datasets (NSL-KDD) and catalonia independence corpus intrusion detection system (CIC-IDS-2017) datasets, achieving an average F1-score of 96% for binary classification on NSL-KDD and 85% for binary classification on CIC-IDS-2017, while for multi-classification, the proposed model achieved an average F1-score of 82% and 96% for NSL-KDD and CIC-IDS-2017 successively.